Define the claim
Write the claim at the right scale: a particular word is borrowed; a particular construction may have diffused; a script was adapted; an institution issued a text in multiple languages; or a community shifted domains. These are not interchangeable. “The languages influenced one another” is too vague to test. A stronger claim names the form, the proposed direction and the dated period, and separates observations from interpretations.
State the null and alternatives. For a word, consider inheritance, chance and an intermediate donor. For a syntactic pattern, consider parallel grammaticalization and genre conventions. For script history, consider independent form development and editorial normalization. For language shift, consider missing documentation and learned continuity. Each alternative should imply some observation that could change your assessment.
Build the evidence chain
Start with the object or earliest edition and record the visible text, uncertainty marks and location. List later witnesses without mixing composition date and copy date. Gather each language’s internal history and the relevant neighboring languages. Compare forms and meanings in context; then test chronology and possible social routes. If a proposed pathway requires an unattested intermediary, name it as an inference. Retain the exact source and page or corpus identifier behind every decisive step.
For lexical work, WOLD’s distinction between immediate and earlier donor languages, and its graded borrowed-status field, provide a useful audit. For multilingual texts, the British Museum, UNESCO and Oracc resources show three different witness types: a royal decree, an imperial monument and a scribal teaching tradition. They cannot all support the same kind of sociolinguistic conclusion.
Write the result
A responsible result has three layers. First, describe what is directly attested: which signs, forms, date and physical context. Second, give the analytical comparison: normalizations, proposed cognates or correspondences, and translation decisions. Third, state the historical inference and its degree of confidence. Mark reconstructions and missing data explicitly. If the two best analyses remain tied, say so rather than forcing a single narrative.
Source exercise: use the Rosetta Stone and the Bisotun inscription to test the claim “parallel inscriptions prove people were bilingual.” Both monuments demonstrate multilingual publication. They do not directly count bilingual speakers. A refined claim asks which institutions commissioned the versions, who prepared them, and what other evidence bears on audiences and language competence.
Modern-language examples on this page clarify a linguistic pattern. They do not establish unattested sounds or forms for an ancient language. Follow the cited sources below and keep editorial readings separate from the surviving trace.